Main Analysis

The following barcharts plot means with 95% confidence intervals. This first figure plots the overall rate of belief, click and sharing outcomes for stimuli (true and false).

Tables are the results of mixed effect models run as specified in the pre-analysis plan with linear regressions. We use mixed effect models because we are not interested in the “effect” of each specific respondent/headline on the outcome (belief, click, share), rather we hope our results generalize to other headlines and respondents. However, we know that the outcomes from the same respondent and for the same headline are likely correlated, so we want to control for that by including headline and respondent random effects.

For robustness, we also present results using logit models. These tables present odds ratios (odds ratios > 1 indicate a positive relationship, < 1 indicates a negative relationship). In all models, the emotion and veracity (true/false) variables have been recoded such that true = .5, false = -.5 and emotion = .5, absence of that emotion = -.5.

Neutral or any emotion

Linear regression mixed effects model


	Dependent variable:

	belief	click	share
	(1)	(2)	(3)

neutralC	-0.172^***	-0.206^***	-0.224^***
	(0.029)	(0.019)	(0.019)

trueC	0.263^***	0.071	0.148^*
	(0.090)	(0.066)	(0.079)

neutralC:trueC	0.112^*	-0.032	0.026
	(0.057)	(0.037)	(0.035)

Constant	0.699^***	0.719^***	0.580^***
	(0.045)	(0.033)	(0.040)


Observations	13,410	13,409	13,409
Log Likelihood	-5,996.359	-5,329.677	-6,794.762
Akaike Inf. Crit.	12,010.720	10,677.350	13,607.520
Bayesian Inf. Crit.	12,078.250	10,744.890	13,675.060

Note:	p<0.1; p<0.05; p<0.01

Logistic mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	4.21	<0.001	4.92	<0.001	1.96	0.010
neutralC	0.30	<0.001	0.22	<0.001	0.23	<0.001
trueC	7.54	<0.001	1.96	0.121	2.96	0.031
neutralC * trueC	0.85	0.541	0.51	0.005	0.92	0.695
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.44 _respID		2.37 _respID		3.48 _respID
	0.77 _headline		0.46 _headline		0.63 _headline
τ₁₁	0.11 _{headline.neutralC}		0.10 _{headline.neutralC}		0.05 _{headline.neutralC}
ρ₀₁	0.01 _headline		-0.68 _headline		-0.80 _headline
ICC	0.40		0.47		0.56
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.186 / 0.515		0.070 / 0.510		0.068 / 0.592

Anger

Linear regression mixed effects model


	Dependent variable:

	belief	click	share
	(1)	(2)	(3)

angerC	-0.121^***	-0.097^***	-0.135^***
	(0.046)	(0.036)	(0.039)

trueC	0.270^***	0.129^**	0.181^**
	(0.080)	(0.059)	(0.072)

angerC:trueC	0.110	0.089	0.088
	(0.091)	(0.072)	(0.079)

Constant	0.714^***	0.760^***	0.613^***
	(0.040)	(0.030)	(0.037)


Observations	13,410	13,409	13,409
Log Likelihood	-6,083.703	-5,456.297	-6,890.745
Akaike Inf. Crit.	12,185.410	10,930.590	13,799.490
Bayesian Inf. Crit.	12,252.940	10,998.130	13,867.020

Note:	p<0.1; p<0.05; p<0.01

Logistic regression mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	4.59	<0.001	6.49	<0.001	2.42	<0.001
angerC	0.41	<0.001	0.49	0.001	0.40	<0.001
trueC	8.21	<0.001	2.87	0.015	3.55	0.007
angerC * trueC	1.00	0.995	1.38	0.456	1.36	0.512
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.48 _respID		2.52 _respID		3.70 _respID
	0.67 _headline		0.45 _headline		0.53 _headline
τ₁₁	0.47 _{headline.angerC}		0.36 _{headline.angerC}		0.45 _{headline.angerC}
ρ₀₁	-0.12 _headline		-0.12 _headline		-0.22 _headline
ICC	0.41		0.48		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.183 / 0.521		0.046 / 0.508		0.058 / 0.598

Fear

Linear regression mixed effects model


	Dependent variable:

	belief	click	share
	(1)	(2)	(3)

fearC	0.073^**	0.078^***	0.099^***
	(0.031)	(0.022)	(0.034)

trueC	0.226^***	0.090^*	0.163^***
	(0.058)	(0.048)	(0.061)

fearC:trueC	0.016	0.024	0.077
	(0.061)	(0.045)	(0.067)

Constant	0.793^***	0.826^***	0.700^***
	(0.029)	(0.025)	(0.031)


Observations	13,410	13,409	13,409
Log Likelihood	-6,121.023	-5,483.299	-6,906.449
Akaike Inf. Crit.	12,260.050	10,984.600	13,830.900
Bayesian Inf. Crit.	12,327.580	11,052.130	13,898.430

Note:	p<0.1; p<0.05; p<0.01

Logistic regression mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	8.83	<0.001	11.23	<0.001	4.44	<0.001
fearC	2.26	<0.001	2.24	<0.001	2.13	<0.001
trueC	10.08	<0.001	2.85	0.011	3.81	0.002
fearC * trueC	2.21	0.016	1.86	0.063	2.17	0.035
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.51 _respID		2.53 _respID		3.69 _respID
	0.55 _headline		0.39 _headline		0.43 _headline
τ₁₁	0.15 _{headline.fearC}		0.16 _{headline.fearC}		0.25 _{headline.fearC}
ρ₀₁	-0.38 _headline		-0.25 _headline		-0.59 _headline
ICC	0.40		0.48		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.196 / 0.516		0.054 / 0.505		0.060 / 0.594

Happiness

Linear regression mixed effects model


	Dependent variable:

	belief	click	share
	belief	click	share
	(1)	(2)	(3)

happyC	0.155^***	0.108^***	0.158^***
	(0.031)	(0.025)	(0.026)

trueC	0.191^***	0.077	0.109^*
	(0.057)	(0.049)	(0.062)

happyC:trueC	-0.196^***	-0.088^*	-0.189^***
	(0.062)	(0.049)	(0.052)

Constant	0.802^***	0.826^***	0.705^***
	(0.029)	(0.025)	(0.032)


Observations	13,410	13,409	13,409
Log Likelihood	-5,853.160	-5,393.049	-6,715.281
Akaike Inf. Crit.	11,724.320	10,804.100	13,448.560
Bayesian Inf. Crit.	11,791.850	10,871.630	13,516.090

Note:	p<0.1; p<0.05; p<0.01

Logistic mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	9.15	<0.001	12.28	<0.001	5.11	<0.001
happyC	3.21	<0.001	3.19	<0.001	3.71	<0.001
trueC	7.07	<0.001	2.25	0.066	2.54	0.057
happyC * trueC	0.48	0.031	0.52	0.121	0.32	0.001
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.42 _respID		2.55 _respID		3.90 _respID
	0.57 _headline		0.45 _headline		0.58 _headline
τ₁₁	0.17 _{headline.happyC}		0.32 _{headline.happyC}		0.17 _{headline.happyC}
ρ₀₁	-0.70 _headline		0.04 _headline		-0.10 _headline
ICC	0.40		0.48		0.58
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.195 / 0.514		0.066 / 0.517		0.082 / 0.614

Sadness

Linear regression mixed effects model


	Dependent variable:

	belief	click	share
	belief	click	share
	(1)	(2)	(3)

sadC	0.020	0.018	0.002
	(0.024)	(0.025)	(0.029)

trueC	0.232^***	0.105^*	0.173^**
	(0.072)	(0.056)	(0.073)

sadC:trueC	0.042	0.057	0.092
	(0.048)	(0.051)	(0.057)

Constant	0.766^***	0.797^***	0.659^***
	(0.036)	(0.029)	(0.037)


Observations	13,410	13,409	13,409
Log Likelihood	-6,133.851	-5,493.764	-6,947.877
Akaike Inf. Crit.	12,285.700	11,005.530	13,913.750
Bayesian Inf. Crit.	12,353.240	11,073.060	13,981.290

Note:	p<0.1; p<0.05; p<0.01

Logistic mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	7.00	<0.001	8.54	<0.001	3.28	<0.001
sadC	1.53	0.007	1.28	0.177	1.04	0.803
trueC	9.56	<0.001	2.98	0.016	3.81	0.007
sadC * trueC	2.49	0.004	2.12	0.038	2.03	0.025
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.48 _respID		2.49 _respID		3.68 _respID
	0.80 _headline		0.50 _headline		0.61 _headline
τ₁₁	0.18 _{headline.sadC}		0.26 _{headline.sadC}		0.19 _{headline.sadC}
ρ₀₁	0.47 _headline		0.08 _headline		-0.26 _headline
ICC	0.40		0.48		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.182 / 0.513		0.042 / 0.501		0.045 / 0.591

Surprise

Linear regression mixed effects model

Model with only random intercepts due to singularity.


	Dependent variable:

	belief	click	share
	belief	click	share
	(1)	(2)	(3)

surpriseC	0.016^**	0.090^***	0.061^***
	(0.008)	(0.007)	(0.008)

trueC	0.214^***	0.098	0.144^*
	(0.077)	(0.067)	(0.086)

surpriseC:trueC	-0.052^***	-0.022	-0.038^**
	(0.015)	(0.014)	(0.016)

Constant	0.767^***	0.817^***	0.678^***
	(0.039)	(0.034)	(0.043)


Observations	13,410	13,409	13,409
Log Likelihood	-6,144.651	-5,439.658	-6,937.499
Akaike Inf. Crit.	12,303.300	10,893.320	13,889.000
Bayesian Inf. Crit.	12,355.830	10,945.840	13,941.520

Note:	p<0.1; p<0.05; p<0.01

Logit mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	6.63	<0.001	10.25	<0.001	3.73	<0.001
surpriseC	1.07	0.286	2.26	<0.001	1.50	<0.001
trueC	7.62	<0.001	2.88	0.038	3.23	0.036
surpriseC * trueC	0.78	0.044	1.15	0.289	0.94	0.599
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.48 _respID		2.37 _respID		3.54 _respID
	0.79 _headline		0.64 _headline		0.78 _headline
ICC	0.41		0.48		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.162 / 0.504		0.050 / 0.504		0.044 / 0.586

Disgust

Linear regression mixed effects model

Only random slope for headline, otherwise singular for belief logit model.


	Dependent variable:

	belief	click	share
	belief	click	share
	(1)	(2)	(3)

disgustC	-0.137^***	-0.147^***	-0.168^***
	(0.042)	(0.034)	(0.039)

trueC	0.250^***	0.132^**	0.177^**
	(0.077)	(0.058)	(0.077)

disgustC:trueC	0.081	0.110	0.086
	(0.084)	(0.068)	(0.078)

Constant	0.706^***	0.739^***	0.599^***
	(0.039)	(0.029)	(0.039)


Observations	13,410	13,409	13,409
Log Likelihood	-6,091.117	-5,417.688	-6,869.702
Akaike Inf. Crit.	12,200.230	10,853.380	13,757.400
Bayesian Inf. Crit.	12,267.770	10,920.910	13,824.940

Note:	p<0.1; p<0.05; p<0.01

Logit mixed effects model

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	4.27	<0.001	5.72	<0.001	2.23	0.002
disgustC	0.36	<0.001	0.35	<0.001	0.34	<0.001
trueC	6.94	<0.001	2.88	0.015	3.50	0.014
disgustC * trueC	0.79	0.633	1.59	0.269	1.42	0.440
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.51 _respID		2.60 _respID		3.70 _respID
	0.59 _headline		0.44 _headline		0.63 _headline
τ₁₁	0.44 _{headline.disgustC}		0.29 _{headline.disgustC}		0.39 _{headline.disgustC}
ρ₀₁	-0.24 _headline		-0.03 _headline		0.10 _headline
ICC	0.41		0.49		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.176 / 0.516		0.052 / 0.514		0.059 / 0.597

Anger, Fear, Sadness, Happiness and Belief

Linear mixed effects models with emotion strength


	Dependent variable:

	belief
	(1)	(2)	(3)	(4)	(5)	(6)

emostrengthanger2	-0.021^**
	(0.009)

emostrengthfear2		0.012^**
		(0.006)

emostrengthsad2			0.004
			(0.004)

emostrengthhappy2				0.028^***
				(0.005)

emostrengthsurprise2					0.005^***
					(0.001)

emostrengthdisgust2						-0.022^***
						(0.008)

trueC	0.217^***	0.218^***	0.212^***	0.287^***	0.242^***	0.211^***
	(0.080)	(0.079)	(0.079)	(0.084)	(0.077)	(0.078)

emostrengthanger2:trueC	0.021
	(0.017)

emostrengthfear2:trueC		0.004
		(0.011)

emostrengthsad2:trueC			0.008
			(0.009)

emostrengthhappy2:trueC				-0.032^***
				(0.010)

emostrengthsurprise2:trueC					-0.010^***
					(0.003)

emostrengthdisgust2:trueC						0.017
						(0.017)

Constant	0.772^***	0.757^***	0.756^***	0.725^***	0.756^***	0.772^***
	(0.040)	(0.039)	(0.040)	(0.042)	(0.039)	(0.039)


Observations	13,410	13,410	13,410	13,410	13,410	13,410
Log Likelihood	-6,094.351	-6,124.775	-6,136.415	-5,846.123	-6,141.368	-6,103.722
Akaike Inf. Crit.	12,206.700	12,267.550	12,290.830	11,710.250	12,296.740	12,225.440
Bayesian Inf. Crit.	12,274.240	12,335.080	12,358.360	11,777.780	12,349.260	12,292.980

Note:	p<0.1; p<0.05; p<0.01

Main results interacted with partisanship

There is no significant interaction between identifying with a party “party_close” and veracity or the triple interaction with the emotion for models where the outcome is belief. For models predicting intention to share in the table below, the triple interaction is not significant but veracity[true] x party_close is greater than 1 and statistically significant in the happy and fear models, indicating that individuals who identify with a party are more likely to share true than false info, compare to people who do not identify with any party. Overall, partisans are more likely to share stimuli than non-partisans. [below are logit models]

	share		share		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	4.48	<0.001	2.74	<0.001	3.62	<0.001
happyC	4.24	<0.001
trueC	2.53	0.058	3.66	0.010	3.79	0.002
party_close	1.84	<0.001	2.30	<0.001	2.53	<0.001
happyC * trueC	0.33	0.002
happyC * party_close	0.52	<0.001
trueC * party_close	1.12	0.482	1.38	0.037	1.11	0.585
(happyC * trueC) * party_close	0.72	0.335
sadC			1.04	0.789
sadC * trueC			1.99	0.031
sadC * party_close			0.95	0.747
(sadC * trueC) * party_close			1.16	0.628
fearC					2.02	<0.001
fearC * trueC					2.33	0.026
fearC * party_close					1.25	0.273
(fearC * trueC) * party_close					0.68	0.325
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	3.74 _respID		3.51 _respID		3.54 _respID
	0.57 _headline		0.61 _headline		0.44 _headline
τ₁₁	0.18 _{headline.happyC}		0.18 _{headline.sadC}		0.25 _{headline.fearC}
ρ₀₁	-0.12 _headline		-0.29 _headline		-0.57 _headline
ICC	0.57		0.56		0.56
N	1320 _respID		1320 _respID		1320 _respID
	10 _headline		10 _headline		10 _headline
Observations	13200		13200		13200
Marginal R² / Conditional R²	0.099 / 0.614		0.064 / 0.590		0.079 / 0.594

The below table interacts main models with APC (ruling party partisanship) - among individuals who say they identify with a political party. A majority of respondents (60%) do not say they feel close to any political party, which are used in the table above. [below are logit models w/ only random intercepts]

	belief		belief		belief
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	10.98	<0.001	9.70	<0.001	12.00	<0.001
happyC	2.54	<0.001
trueC	9.00	<0.001	13.34	<0.001	12.63	<0.001
APC	1.54	0.097	1.01	0.980	1.16	0.587
happyC * trueC	0.28	0.009
happyC * APC	1.96	0.097
trueC * APC	0.78	0.521	0.59	0.081	0.72	0.423
(happyC * trueC) * APC	3.27	0.135
sadC			1.49	0.082
sadC * trueC			1.48	0.378
sadC * APC			0.68	0.226
(sadC * trueC) * APC			1.89	0.299
fearC					2.38	0.003
fearC * trueC					1.24	0.704
fearC * APC					1.02	0.959
(fearC * trueC) * APC					3.03	0.182
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	2.01 _respID		2.19 _respID		2.22 _respID
	0.64 _headline		0.62 _headline		0.58 _headline
ICC	0.45		0.46		0.46
N	306 _respID		306 _respID		306 _respID
	10 _headline		10 _headline		10 _headline
Observations	3060		3060		3060
Marginal R² / Conditional R²	0.206 / 0.561		0.179 / 0.558		0.202 / 0.570

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	8.87	<0.001	11.79	<0.001	4.92	<0.001
trueC	12.97	<0.001	3.39	0.004	3.82	0.006
APC	1.17	0.468	2.16	0.007	1.90	0.011
trueC * APC	0.47	0.001	1.16	0.561	0.96	0.841
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	2.20 _respID		3.78 _respID		3.70 _respID
	0.68 _headline		0.38 _headline		0.54 _headline
ICC	0.47		0.56		0.56
N	306 _respID		306 _respID		306 _respID
	10 _headline		10 _headline		10 _headline
Observations	3060		3060		3060
Marginal R² / Conditional R²	0.170 / 0.557		0.071 / 0.589		0.067 / 0.592

Main linear models with controls

Main results for neutral, fear, sadness, and happiness using linear mixed effect regressions - controlling for gender, age, and education level. Belief regression controlling for gender, age, education level.[All models have both headline and respondent random intercepts and random slopes for the emotion wrt headline.]


	Dependent variable:

	belief
	(1)	(2)	(3)	(4)	(5)	(6)	(7)

neutralC	-0.174^***
	(0.029)

fearC		0.074^**
		(0.031)

sadC			0.020
			(0.024)

angerC				-0.120^***
				(0.047)

happyC					0.152^***
					(0.032)

surpriseC						0.017^**
						(0.008)

disgustC							-0.139^***
							(0.043)

trueC	0.271^***	0.228^***	0.234^***	0.274^***	0.192^***	0.215^***	0.251^***
	(0.090)	(0.058)	(0.071)	(0.080)	(0.057)	(0.077)	(0.076)

age	-0.001	-0.001	-0.001	-0.0004	-0.0005	-0.001	-0.0004
	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)

gender	-0.010	0.00004	0.001	0.001	-0.001	-0.001	0.003
	(0.011)	(0.011)	(0.011)	(0.011)	(0.011)	(0.011)	(0.011)

edu	-0.002	-0.002	-0.002	-0.002	-0.002	-0.002	-0.002
	(0.005)	(0.005)	(0.005)	(0.005)	(0.005)	(0.005)	(0.005)

neutralC:trueC	0.125^**
	(0.057)

fearC:trueC		0.014
		(0.063)

sadC:trueC			0.041
			(0.048)

angerC:trueC				0.113
				(0.093)

happyC:trueC					-0.200^***
					(0.063)

surpriseC:trueC						-0.055^***
						(0.015)

disgustC:trueC							0.078
							(0.086)

Constant	0.744^***	0.829^***	0.797^***	0.743^***	0.835^***	0.798^***	0.728^***
	(0.058)	(0.048)	(0.052)	(0.055)	(0.046)	(0.054)	(0.054)


Observations	13,190	13,190	13,190	13,190	13,190	13,190	13,190
Log Likelihood	-5,884.218	-6,011.762	-6,026.376	-5,977.404	-5,746.957	-6,035.761	-5,983.126
Akaike Inf. Crit.	11,792.430	12,047.520	12,076.750	11,978.810	11,517.910	12,091.520	11,990.250
Bayesian Inf. Crit.	11,882.280	12,137.370	12,166.600	12,068.650	11,607.760	12,166.390	12,080.100

Note:	p<0.1; p<0.05; p<0.01

Leave one out analysis

The below graphs plot the coefficients on the interaction terms (emotion x veracity) for the linear mixed effect models predicting belief. These graphs plot estimates on the y axis. The points represent the iterated estimates when 1 out of the 10 headlines is removed from the analysis (to ensure that the results are not being driven by a few headlines). The x-axis indicates which headline is left out.

All emotions in single model

Linear regression model:


	Dependent variable:

	belief	click	share
	(1)	(2)	(3)

trueC	0.210^***	0.076^*	0.126^**
	(0.065)	(0.044)	(0.056)

angerC	-0.020^*	-0.013	-0.035^***
	(0.010)	(0.010)	(0.011)

disgustC	-0.032^***	-0.070^***	-0.071^***
	(0.011)	(0.010)	(0.012)

fearC	0.101^***	0.110^***	0.150^***
	(0.010)	(0.009)	(0.010)

happyC	0.228^***	0.179^***	0.245^***
	(0.010)	(0.010)	(0.011)

sadC	0.094^***	0.100^***	0.103^***
	(0.009)	(0.008)	(0.009)

surpriseC	0.063^***	0.128^***	0.110^***
	(0.008)	(0.008)	(0.009)

Constant	0.886^***	0.911^***	0.801^***
	(0.034)	(0.025)	(0.031)


Observations	13,410	13,409	13,409
Log Likelihood	-5,836.878	-5,154.609	-6,534.496
Akaike Inf. Crit.	11,695.760	10,331.220	13,090.990
Bayesian Inf. Crit.	11,778.300	10,413.760	13,173.530

Note:	p<0.1; p<0.05; p<0.01

Secondary Analysis

Are clicking, believing, and sharing behaviors related?

Correlations between outcomes

##           belief     click     share
## belief 1.0000000 0.3813965 0.4558099
## click  0.3813965 1.0000000 0.5557195
## share  0.4558099 0.5557195 1.0000000

Belief and clicking to read more, if possible, are both positively associated with sharing the stimuli, regardless of whether the stimuli was true or false. [Linear mixed effects models]

	share		share		share		share		share
Predictors	Estimates	p	Estimates	p	Estimates	p	Estimates	p	Estimates	p
(Intercept)	0.11	<0.001	0.37	<0.001	0.25	<0.001	0.13	<0.001	0.08	0.005
belief	0.27	<0.001	0.39	<0.001			0.24	<0.001	0.30	<0.001
trueC	0.06	0.174	0.09	0.152	0.11	0.035
click	0.44	<0.001			0.52	<0.001	0.47	<0.001	0.42	<0.001
belief * trueC	-0.06	<0.001	-0.05	0.004
trueC * click	0.04	0.017
click * trueC					-0.02	0.158
Random Effects
σ²	0.10		0.12		0.11		0.09		0.11
τ₀₀	0.03 _respID		0.05 _respID		0.04 _respID		0.03 _respID		0.04 _respID
	0.00 _headline		0.01 _headline		0.01 _headline		0.01 _headline		0.00 _headline
ICC	0.27		0.34		0.28		0.28		0.29
N	1341 _respID		1341 _respID		1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline		5 _headline		5 _headline
Observations	13409		13409		13409		6704		6705
Marginal R² / Conditional R²	0.309 / 0.499		0.153 / 0.438		0.239 / 0.456		0.257 / 0.468		0.328 / 0.522

Encouragingly, only a small percent of the time do individuals say they want to share a stimuli that they do not believe is true. round(dim(ng.long[ng.long$belief==0 & ng.long$share==1,])[1]/dim(ng.long)[1]*100,digits=1)% of cases do respondents say they want to share something that they do not believe is true. However, respondents are much more willing to share something without clicking to read the full news story. In round(dim(ng.long[ng.long$click==0 & ng.long$share==1,])[1]/dim(ng.long)[1]*100,digits=1)% of cases respondents say they would share the stimuli without clicking to read the full article.

Emotions and time spent on belief question

with emotion	absence emo.	p.value	test
11.75	9.38	0.00	neutral
8.89	9.90	0.00	sad
8.65	9.84	0.00	fear
9.03	9.81	0.02	happy
10.46	9.52	0.27	anger
9.84	9.57	0.50	surprise
9.66	9.65	0.99	disgust

Emotion Regulation

Reappraisal

Logit
	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	2.27	0.047	1.24	0.613	0.38	0.044
erq_reapp	1.20	<0.001	1.40	<0.001	1.47	<0.001
trueC	38.42	<0.001	3.82	0.029	10.03	<0.001
erq_reapp * trueC	0.77	<0.001	0.94	0.294	0.82	0.001
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.49 _respID		2.28 _respID		3.32 _respID
	0.78 _headline		0.64 _headline		0.75 _headline
ICC	0.41		0.47		0.55
N	1228 _respID		1228 _respID		1228 _respID
	10 _headline		10 _headline		10 _headline
Observations	12280		12280		12280
Marginal R² / Conditional R²	0.172 / 0.510		0.049 / 0.496		0.058 / 0.579

Linear
	belief		click		share
Predictors	Estimates	p	Estimates	p	Estimates	p
(Intercept)	0.60	<0.001	0.59	<0.001	0.39	<0.001
erq_reapp	0.03	<0.001	0.04	<0.001	0.05	<0.001
trueC	0.53	<0.001	0.22	0.007	0.39	<0.001
erq_reapp * trueC	-0.05	<0.001	-0.02	0.003	-0.04	<0.001
Random Effects
σ²	0.13		0.12		0.14
τ₀₀	0.02 _respID		0.03 _respID		0.06 _respID
	0.01 _headline		0.01 _headline		0.02 _headline
ICC	0.22		0.27		0.36
N	1228 _respID		1228 _respID		1228 _respID
	10 _headline		10 _headline		10 _headline
Observations	12280		12280		12280
Marginal R² / Conditional R²	0.077 / 0.283		0.020 / 0.283		0.034 / 0.377

Suppression

Logit
	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	4.31	<0.001	4.99	<0.001	0.84	0.615
erq_supp	1.10	0.002	1.13	0.001	1.37	<0.001
trueC	19.49	<0.001	3.37	0.023	4.67	0.007
erq_supp * trueC	0.83	<0.001	0.95	0.172	0.92	0.021
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.50 _respID		2.38 _respID		3.29 _respID
	0.79 _headline		0.64 _headline		0.75 _headline
ICC	0.41		0.48		0.55
N	1248 _respID		1248 _respID		1248 _respID
	10 _headline		10 _headline		10 _headline
Observations	12480		12480		12480
Marginal R² / Conditional R²	0.173 / 0.512		0.041 / 0.500		0.067 / 0.582

Linear
	belief		click		share
Predictors	Estimates	p	Estimates	p	Estimates	p
(Intercept)	0.70	<0.001	0.75	<0.001	0.49	<0.001
erq_supp	0.02	<0.001	0.01	0.007	0.04	<0.001
trueC	0.37	<0.001	0.14	0.045	0.26	0.004
erq_supp * trueC	-0.03	<0.001	-0.01	0.015	-0.02	<0.001
Random Effects
σ²	0.13		0.12		0.14
τ₀₀	0.02 _respID		0.03 _respID		0.06 _respID
	0.01 _headline		0.01 _headline		0.02 _headline
ICC	0.22		0.27		0.35
N	1248 _respID		1248 _respID		1248 _respID
	10 _headline		10 _headline		10 _headline
Observations	12480		12480		12480
Marginal R² / Conditional R²	0.076 / 0.283		0.015 / 0.284		0.039 / 0.379

Cognitive Reflection

% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu % Date and time: Thu, Dec 17, 2020 - 14:46:07

Correlations with CRT


	Dependent variable:

	crt_index
	(1)	(2)	(3)	(4)	(5)

erq_reapp	-0.243^***
	(0.054)

erq_supp	-0.156^***
	(0.034)

age	-0.036^***	-0.035^***	-0.035^***	-0.035^***	-0.036^***
	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)

factor(gender)1	0.247^**	0.174^*	0.183^*	0.158	0.187^*
	(0.100)	(0.100)	(0.100)	(0.099)	(0.100)

edu	0.065	0.071^*	0.070^*	0.075^*	0.070^*
	(0.041)	(0.042)	(0.042)	(0.041)	(0.042)

factor(urban)1	0.231^**	0.219^*	0.222^**	0.215^*	0.238^**
	(0.111)	(0.112)	(0.112)	(0.112)	(0.112)

factor(voted)1	-0.205^**	-0.309^***	-0.304^***	-0.319^***	-0.320^***
	(0.094)	(0.093)	(0.093)	(0.093)	(0.093)

christ	0.178^*	0.155	-0.167		0.070
	(0.105)	(0.105)	(0.325)		(0.104)

religiosity	-0.058	-0.062	-0.135	-0.058
	(0.045)	(0.045)	(0.082)	(0.045)

christ:religiosity			0.103
			(0.098)

Constant	4.297^***	2.235^***	2.464^***	2.318^***	2.122^***
	(0.464)	(0.365)	(0.426)	(0.361)	(0.338)


Observations	1,172	1,193	1,193	1,193	1,210
R²	0.082	0.040	0.041	0.038	0.039
Adjusted R²	0.075	0.034	0.034	0.033	0.034
Residual Std. Error	1.500 (df = 1162)	1.529 (df = 1185)	1.529 (df = 1184)	1.530 (df = 1186)	1.543 (df = 1203)
F Statistic	11.596^*** (df = 9; 1162)	7.027^*** (df = 7; 1185)	6.286^*** (df = 8; 1184)	7.831^*** (df = 6; 1186)	8.078^*** (df = 6; 1203)

Note:	p<0.1; p<0.05; p<0.01

Correlations with discernment


	Dependent variable:

	discern
	(1)	(2)	(3)	(4)	(5)	(6)	(7)

erq_reapp	-0.274^***
	(0.054)

erq_supp		-0.181^***
		(0.033)

crt_index			0.192^***	0.191^***			0.193^***
			(0.029)	(0.029)			(0.029)

age	0.013^*	0.009	0.018^**	0.018^**	0.012^*	0.012^*	0.018^**
	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)

factor(gender)1	0.153	0.252^**	0.122	0.134	0.155	0.164^*	0.121
	(0.102)	(0.102)	(0.102)	(0.101)	(0.098)	(0.097)	(0.102)

edu	0.061	0.055	0.049	0.046	0.052	0.050	0.047
	(0.042)	(0.042)	(0.042)	(0.042)	(0.040)	(0.040)	(0.042)

factor(urban)1	0.269^**	0.279^**	0.212^*	0.216^*	0.267^**	0.269^**	0.223^**
	(0.113)	(0.112)	(0.113)	(0.113)	(0.109)	(0.109)	(0.113)

factor(APC_party)1	-0.376^**	-0.370^**	-0.363^**	-0.344^**	-0.395^***	-0.383^***
	(0.146)	(0.144)	(0.144)	(0.143)	(0.142)	(0.141)

factor(party_close)1							-0.120
							(0.108)

factor(voted)1	0.043	0.049	0.026	0.031	-0.017	-0.013	0.007
	(0.097)	(0.097)	(0.096)	(0.096)	(0.094)	(0.094)	(0.097)

factor(christ)1	-0.027	-0.121	-0.121		-0.084		-0.088
	(0.108)	(0.106)	(0.106)		(0.105)		(0.106)

religiosity	-0.067	-0.101^**	-0.081^*	-0.085^*	-0.097^**	-0.099^**	-0.082^*
	(0.045)	(0.045)	(0.045)	(0.045)	(0.044)	(0.044)	(0.045)

someuse	0.039	0.030	0.040^*	0.040	0.036	0.036	0.040
	(0.025)	(0.024)	(0.025)	(0.025)	(0.023)	(0.023)	(0.025)

Constant	6.308^***	5.886^***	4.476^***	4.416^***	5.012^***	4.962^***	4.469^***
	(0.516)	(0.469)	(0.447)	(0.444)	(0.427)	(0.422)	(0.448)


Observations	1,192	1,211	1,193	1,193	1,276	1,276	1,193
R²	0.048	0.051	0.062	0.061	0.028	0.028	0.058
Adjusted R²	0.040	0.043	0.054	0.054	0.021	0.021	0.050
Residual Std. Error	1.542 (df = 1181)	1.538 (df = 1200)	1.532 (df = 1182)	1.532 (df = 1183)	1.545 (df = 1266)	1.544 (df = 1267)	1.535 (df = 1182)
F Statistic	5.922^*** (df = 10; 1181)	6.405^*** (df = 10; 1200)	7.833^*** (df = 10; 1182)	8.558^*** (df = 9; 1183)	4.057^*** (df = 9; 1266)	4.486^*** (df = 8; 1267)	7.290^*** (df = 10; 1182)

Note:	p<0.1; p<0.05; p<0.01

Correlates of Discernment

Are emotion regulation and suppression correlated with felt emotions?


	Dependent variable:

	emo.index
	(1)	(2)

erq_reapp	-0.095
	(0.061)

erq_supp		-0.055
		(0.038)

age	-0.015^*	-0.016^**
	(0.008)	(0.008)

factor(gender)1	-0.405^***	-0.380^***
	(0.114)	(0.114)

edu	0.016	0.019
	(0.048)	(0.047)

factor(urban)1	0.031	0.045
	(0.128)	(0.127)

factor(voted)1	-0.494^***	-0.468^***
	(0.107)	(0.107)

factor(christ)1	0.108	0.097
	(0.122)	(0.120)

religiosity	-0.067	-0.072
	(0.051)	(0.051)

Constant	2.524^***	2.225^***
	(0.522)	(0.456)


Observations	1,192	1,211
R²	0.050	0.048
Adjusted R²	0.044	0.042
Residual Std. Error	1.748 (df = 1183)	1.751 (df = 1202)
F Statistic	7.795^*** (df = 8; 1183)	7.599^*** (df = 8; 1202)

Note:	p<0.1; p<0.05; p<0.01

Does being asked about emotions/belief first matter?

	belief		click		share
Predictors	Odds Ratios	p	Odds Ratios	p	Odds Ratios	p
(Intercept)	2.15	0.057	5.21	<0.001	1.80	0.142
veracity [true]	7.68	<0.001	2.72	0.042	3.04	0.045
A_or_B [B]	1.18	0.059	1.07	0.529	1.19	0.156
veracity [true] * A_or_B [B]	1.09	0.428	0.84	0.083	1.04	0.692
Random Effects
σ²	3.29		3.29		3.29
τ₀₀	1.49 _respID		2.53 _respID		3.66 _respID
	0.80 _headline		0.60 _headline		0.76 _headline
ICC	0.41		0.49		0.57
N	1341 _respID		1341 _respID		1341 _respID
	10 _headline		10 _headline		10 _headline
Observations	13410		13409		13409
Marginal R² / Conditional R²	0.164 / 0.506		0.032 / 0.503		0.041 / 0.591

	belief		click		share
Predictors	Estimates	p	Estimates	p	Estimates	p
(Intercept)	0.64	<0.001	0.84	<0.001	0.70	<0.001
A_or_B [B]	0.03	0.236	0.01	0.502	0.01	0.549
Observations	1341		1341		1341
R² / R² adjusted	0.001 / 0.000		0.000 / -0.000		0.000 / -0.000

Are people who got more headlines WRONG more likely to seek information after debrief?

No, there does not seem to be any strong predictors of clicking for more information at the end of the survey. But overall, very few people did this – 66 people clicked on Africa check links and 34 people clicked on WHO links.


	Dependent variable:

	click_WHO	click_africacheck	click_WHOwhatsapp	click_ACwhatsapp	click.any
	(1)	(2)	(3)	(4)	(5)

emo.index	0.0002	0.001	-0.001	-0.002	-0.003
	(0.001)	(0.001)	(0.003)	(0.003)	(0.004)

corr.index	-0.0004	-0.001	-0.001	-0.001	-0.001
	(0.001)	(0.001)	(0.003)	(0.004)	(0.005)

age	-0.0002	-0.0001	-0.0003	-0.0004	-0.001
	(0.0003)	(0.0004)	(0.001)	(0.001)	(0.001)

factor(gender)1	-0.005	0.009^*	-0.007	0.036^***	0.034^**
	(0.004)	(0.005)	(0.010)	(0.014)	(0.016)

edu	-0.001	-0.002	-0.003	0.008	0.007
	(0.002)	(0.002)	(0.004)	(0.006)	(0.007)

factor(urban)1	0.005	-0.0004	-0.018	-0.029^*	-0.042^**
	(0.004)	(0.006)	(0.011)	(0.015)	(0.018)

factor(voted)1	-0.004	0.008	-0.012	0.001	-0.009
	(0.003)	(0.005)	(0.009)	(0.013)	(0.015)

factor(christ)1	-0.001	0.006	0.002	0.010	0.003
	(0.004)	(0.006)	(0.011)	(0.014)	(0.017)

religiosity	0.001	0.003	0.008^*	0.006	0.014^*
	(0.002)	(0.002)	(0.004)	(0.006)	(0.007)

Constant	0.018	0.009	0.066^*	-0.035	0.028
	(0.015)	(0.021)	(0.040)	(0.054)	(0.064)


Observations	1,219	1,219	1,219	1,219	1,276
R²	0.007	0.009	0.009	0.013	0.014
Adjusted R²	-0.0001	0.002	0.002	0.006	0.007
Residual Std. Error	0.057 (df = 1209)	0.081 (df = 1209)	0.155 (df = 1209)	0.211 (df = 1209)	0.253 (df = 1266)
F Statistic	0.981 (df = 9; 1209)	1.246 (df = 9; 1209)	1.250 (df = 9; 1209)	1.755^* (df = 9; 1209)	1.992^** (df = 9; 1266)

Note:	p<0.1; p<0.05; p<0.01

Are people who felt more emotions (and particular kinds) more likely to seek information?


	Dependent variable:

	click_WHO	click_africacheck	click_WHOwhatsapp	click_ACwhatsapp	click.any
	(1)	(2)	(3)	(4)	(5)

anger.ind	0.002	-0.001	0.005	-0.001	0.003
	(0.001)	(0.002)	(0.003)	(0.004)	(0.005)

fear.ind	-0.001	-0.003	0.002	-0.005	-0.006
	(0.001)	(0.002)	(0.003)	(0.005)	(0.005)

sad.ind	0.002	-0.0005	-0.004	0.002	0.001
	(0.001)	(0.001)	(0.003)	(0.004)	(0.004)

happy.ind	-0.0003	-0.002	0.002	0.0001	0.002
	(0.001)	(0.002)	(0.003)	(0.004)	(0.005)

surprise.ind	-0.0004	0.001	-0.001	0.009^***	0.007^*
	(0.001)	(0.001)	(0.002)	(0.003)	(0.004)

disgust.ind	-0.002	0.005^**	-0.0001	-0.0004	0.001
	(0.001)	(0.002)	(0.004)	(0.005)	(0.006)

corr.index	-0.001	-0.002	-0.001	0.001	-0.0003
	(0.001)	(0.002)	(0.003)	(0.004)	(0.005)

age	-0.0002	-0.0001	-0.0003	-0.0005	-0.001
	(0.0003)	(0.0004)	(0.001)	(0.001)	(0.001)

factor(gender)1	-0.004	0.007	-0.006	0.033^**	0.033^**
	(0.004)	(0.005)	(0.010)	(0.014)	(0.016)

edu	-0.001	-0.002	-0.003	0.008	0.007
	(0.002)	(0.002)	(0.004)	(0.006)	(0.007)

factor(urban)1	0.004	0.0001	-0.018	-0.030^*	-0.044^**
	(0.004)	(0.006)	(0.011)	(0.015)	(0.018)

factor(voted)1	-0.004	0.008	-0.012	-0.0002	-0.009
	(0.003)	(0.005)	(0.009)	(0.013)	(0.015)

factor(christ)1	-0.001	0.007	0.003	0.010	0.004
	(0.004)	(0.006)	(0.011)	(0.014)	(0.017)

religiosity	0.001	0.003	0.008^*	0.006	0.014^**
	(0.002)	(0.002)	(0.004)	(0.006)	(0.007)

Constant	0.016	0.014	0.059	-0.059	-0.004
	(0.015)	(0.022)	(0.042)	(0.057)	(0.067)


Observations	1,219	1,219	1,219	1,219	1,276
R²	0.012	0.016	0.014	0.020	0.017
Adjusted R²	0.001	0.005	0.002	0.008	0.006
Residual Std. Error	0.057 (df = 1204)	0.081 (df = 1204)	0.155 (df = 1204)	0.210 (df = 1204)	0.253 (df = 1261)
F Statistic	1.053 (df = 14; 1204)	1.402 (df = 14; 1204)	1.193 (df = 14; 1204)	1.736^** (df = 14; 1204)	1.557^* (df = 14; 1261)

Note:	p<0.1; p<0.05; p<0.01

Headline Descriptives

headline	false_1	false_2	false_3	false_4	false_5	true_1	true_2	true_3	true_4	true_5
fear.mean	0.03	0.15	0.04	0.05	0.09	0.21	0.03	0.39	0.13	0.46
fear.sd	0.18	0.35	0.21	0.21	0.29	0.41	0.18	0.49	0.34	0.50
anger.mean	0.04	0.34	0.05	0.19	0.21	0.04	0.28	0.05	0.06	0.13
anger.sd	0.20	0.47	0.22	0.39	0.41	0.19	0.45	0.21	0.25	0.34
sad.mean	0.05	0.20	0.04	0.10	0.16	0.53	0.13	0.46	0.23	0.51
sad.sd	0.21	0.40	0.20	0.30	0.36	0.50	0.34	0.50	0.42	0.50
happy.mean	0.51	0.04	0.50	0.14	0.09	0.03	0.06	0.14	0.49	0.04
happy.sd	0.50	0.19	0.50	0.35	0.28	0.17	0.24	0.35	0.50	0.19
surprise.mean	0.43	0.29	0.42	0.42	0.28	0.32	0.33	0.12	0.11	0.17
surprise.sd	0.50	0.45	0.49	0.49	0.45	0.47	0.47	0.32	0.31	0.38
disgust.mean	0.04	0.21	0.05	0.17	0.27	0.03	0.28	0.03	0.03	0.05
disgust.sd	0.20	0.41	0.21	0.37	0.44	0.17	0.45	0.16	0.17	0.22
neutral.mean	0.11	0.11	0.11	0.14	0.16	0.10	0.14	0.10	0.14	0.05
neutral.sd	0.31	0.31	0.31	0.35	0.36	0.30	0.35	0.30	0.34	0.21
belief.mean	0.66	0.55	0.68	0.67	0.72	0.94	0.59	0.96	0.94	0.96
belief.sd	0.47	0.50	0.47	0.47	0.45	0.23	0.49	0.20	0.25	0.20
click.mean	0.84	0.73	0.83	0.72	0.65	0.86	0.63	0.92	0.87	0.94
click.sd	0.36	0.44	0.38	0.45	0.48	0.35	0.48	0.28	0.34	0.24
share.mean	0.71	0.51	0.68	0.55	0.53	0.71	0.46	0.86	0.79	0.87
share.sd	0.45	0.50	0.47	0.50	0.50	0.45	0.50	0.35	0.40	0.34

Table presenting the prevalence of felt emotions across all headlines – disaggregated by true/false.

Stimuli

Images of the 10 headlines – 5 true and 5 false and included below.

$False 1$

False 1

False 2

$False 3$

False 3

False 4

False 5

True 1

$True 2$

True 2

True 3

True 4

True 5

Sample Descriptives

**Descriptive statistics of study respondents**

Statistic	N	Mean	St. Dev.	Min	Pctl(25)	Median	Pctl(75)	Max

Female	1,325	0.32	0.47	0.00	0.00	0.00	1.00	1.00
Age	1,321	27.54	7.05	18.00	22.00	26.00	31.00	62.00
Some university	1,321	0.76	0.43	0.00	1.00	1.00	1.00	1.00
Urban	1,297	0.80	0.40	0.00	1.00	1.00	1.00	1.00
Employed	1,316	0.49	0.50	0.00	0.00	0.00	1.00	1.00
Christian	1,316	0.76	0.43	0.00	1.00	1.00	1.00	1.00
Religiosity	1,298	3.24	0.99	0.00	3.00	4.00	4.00	4.00
Voted	1,318	0.45	0.50	0.00	0.00	0.00	1.00	1.00
APC supporter	306	0.47	0.50	0.00	0.00	0.00	1.00	1.00
Daily FB user	1,314	0.74	0.44	0.00	0.00	1.00	1.00	1.00

Nigeria Misinformation: COVID19 Study 1

17 December, 2020

Main Analysis

Neutral or any emotion

Linear regression mixed effects model

Logistic mixed effects model

Anger

Linear regression mixed effects model

Logistic regression mixed effects model

Fear

Linear regression mixed effects model

Logistic regression mixed effects model

Happiness

Linear regression mixed effects model

Logistic mixed effects model

Sadness

Linear regression mixed effects model

Logistic mixed effects model

Surprise

Linear regression mixed effects model

Logit mixed effects model

Disgust

Linear regression mixed effects model

Logit mixed effects model

Anger, Fear, Sadness, Happiness and Belief

Linear mixed effects models with emotion strength

Main results interacted with partisanship

Main linear models with controls

Leave one out analysis

All emotions in single model

Secondary Analysis

Are clicking, believing, and sharing behaviors related?

Emotions and time spent on belief question

Emotion Regulation

Reappraisal

Suppression

Cognitive Reflection

Correlations with CRT

Correlations with discernment

Are emotion regulation and suppression correlated with felt emotions?

Does being asked about emotions/belief first matter?

Are people who got more headlines WRONG more likely to seek information after debrief?

Are people who felt more emotions (and particular kinds) more likely to seek information?

Headline Descriptives

Stimuli

Sample Descriptives